Abstract
Rasch theory of measurement (RTM) was applied to 20 facets present in the English version of the South African Personality Inventory (SAPI-170-E), a culturally relevant personality model for the diverse South African context. Data for analysis (n = 300) was randomly selected from a primary source consisting of almost 4,000 participants used for the validation of the SAPI-188-E, using classical test theory (CTT). CTT, although widely accepted in contemporary scientific research, disregards fundamental assumptions of measurement. RTM has been employed to further increase the potential for reliable and valid personality measures within South Africa’s diverse population. The aim of the analysis making use of RUMM2030 was to identify (1) unidimensionality, (2) threshold discrimination, (3) Rasch model-fit, and (4) facet-targeting, all while keeping psychometric properties stable. Through the identification of these four properties, the reliability and validity of 12 of the 20 SAPI-170-E facets has been identified to have the potential to be improved. It is hoped this study will advise the future editions of the SAPI to be published in South Africa.
M.Phil. (Industrial Psychology)